Omber detection / diagnostic apparatus and method

Abstract

The present invention addresses the problem of specifying the point in time at which wear or notch occurs in a tool of a machine tool. An omen detection / diagnosis device (1) is provided with: a data acquisition unit (10) that acquires time-series data of a load current supplied to a motor of a machine tool (4); a reference data creation unit (12) that calculates, for each time, reference data for kernel density estimation, on the basis of first-lot data divided from time-series data at the time of creation of the reference data; and an inspection execution unit (13) that calculates a kernel density estimation amount for each time point on the basis of second lot data divided from time-series data during inspection and the reference data, and calculates a score indicating the degree of deviation of the second lot data with respect to the first lot data for each time point on the basis of the estimation amount.

Description

Technical Field

[0001] The present invention relates to a prognostic detection / diagnosis device and method for detecting abnormalities in a machine tool. Background Art

[0002] When machining a material using a machine tool, defective products may sometimes be produced due to tool wear or chipping.

[0003] Conventionally, there has been known a technique of measuring the current value of a motor connected to a spindle or the like of a machine tool and evaluating it for the entire batch based on its change to detect tool wear (see Patent Document 1). In addition, there has been known a technique of detecting tool abnormalities for each type of tool using the current value of a motor connected to a spindle or the like of a machine tool and the current value of a motor in charge of tool change (see Patent Document 2).

[0004] The technique disclosed in Patent Document 1 has the following problem: Although it is possible to evaluate the entire batch, in a batch including multiple machinings, the generation position of tool wear or chipping cannot be determined.

[0005] In addition, the technique disclosed in Patent Document 2 has the following problems: In order to divide multiple machining intervals within a batch, an additional measuring device for the motor in charge of tool change is required. Furthermore, since the division of the intervals corresponds to each tool change, there is a problem that continuous different machinings performed on the same tool cannot be separately extracted.

[0006] [Prior Art Documents]

[0007] [Patent Documents]

[0008] [Patent Document 1] Japanese Patent No. 6952318 Gazette

[0009] [Patent Document 2] Japanese Patent No. 6924529 Gazette Summary of the Invention

[0010] [Problems to be Solved by the Invention]

[0011] The present invention has been made to solve the above problems, and an object thereof is to provide a prognostic detection / diagnosis device and method capable of determining the time point at which tool wear or chipping occurs in a machine tool.

[0012] [Technical Means for Solving the Problems]

[0013] The omen detection / diagnosis device of the present invention includes: a data acquisition unit configured to acquire time series data of the load current of a motor supplied to a machine tool; a data storage unit configured to store the time series data; a reference data creation unit configured to calculate reference data for kernel density estimation for each moment based on first batch data segmented from the time series data at the time of creating the reference data; and an inspection execution unit configured to calculate a kernel density estimation amount for each moment based on second batch data segmented from the time series data at the time of inspection and the reference data, and calculate a score indicating the degree of deviation of the second batch data from the first batch data for each moment based on the estimation amount.

[0014] In addition, in a structural example of the omen detection / diagnosis device of the present invention, it further includes a data correction execution unit configured to detect an offset value for a region where the current value in the batch data is below an offset threshold and the state where the current value is below the offset threshold continues for a specified time or more, and subtract or add the offset value from the batch data.

[0015] In addition, in a structural example of the omen detection / diagnosis device of the present invention, the data correction execution unit corrects the time positions of the plurality of first batch data to be consistent, and corrects the second batch data so that the time position is consistent with the first batch data.

[0016] In addition, in a structural example of the omen detection / diagnosis device of the present invention, the reference data creation unit stores in the data storage unit the characteristic quantities of the plurality of first batch data and the reference data in correspondence with the variety data of the workpiece input by the user, and the inspection execution unit determines the variety of the workpiece at the time of inspection based on the characteristic quantities of the first batch data and the second batch data, and calculates the kernel density estimation amount based on the reference data corresponding to the variety of the workpiece at the time of inspection and the second batch data.

[0017] In addition, in a structural example of the omen detection / diagnosis device of the present invention, the characteristic quantity of the first batch data is the average waveform and the average cycle length of the batch data, and the inspection execution unit performs a first variety determination by comparing the average cycle length with the cycle length of the second batch data. When there are two or more variety candidates in the first variety determination, a second variety determination is performed based on the correlation coefficient or cross-correlation between the average waveform and the second batch data. When there are two or more variety candidates in the second variety determination, the kernel density estimation amount is calculated for each remaining variety candidate in the second variety determination, and the variety of the workpiece at the time of inspection is determined based on the kernel density estimation amount.

[0018] Further, in a structural example of the omen detection / diagnosis device of the present invention, the reference data creation unit detects a machining section estimated to have machined a workpiece in the first batch of data, and the inspection execution unit calculates the kernel density estimator only for the section at the same time position as the machining section in the second batch of data.

[0019] Further, in a structural example of the omen detection / diagnosis device of the present invention, the data acquisition unit imports the time series data and acquires the variety data of the workpiece from the machine tool. The reference data creation unit stores the calculated reference data corresponding to the variety data acquired by the data acquisition unit at the time of creating the reference data in the data storage unit. The inspection execution unit acquires the reference data corresponding to the variety data acquired by the data acquisition unit from the data storage unit during inspection, and calculates the kernel density estimator based on the reference data and the second batch of data.

[0020] Further, the omen detection / diagnosis method of the present invention includes: a first step of acquiring time series data of the load current of a motor supplied to a machine tool; a second step of calculating reference data for kernel density estimation for each moment based on a first batch of data segmented from the time series data at the time of creating the reference data; a third step of calculating the kernel density estimator for each moment based on a second batch of data segmented from the time series data at the time of inspection and the reference data; and a fourth step of calculating a score indicating the degree of deviation of the second batch of data from the first batch of data for each moment based on the kernel density estimator.

[0021] [Effects of the Invention]

[0022] According to the present invention, reference data for kernel density estimation is calculated for each moment based on a first batch of data segmented from the time series data at the time of creating the reference data. The kernel density estimator is calculated for each moment based on a second batch of data segmented from the time series data at the time of inspection and the reference data. A score indicating the degree of deviation of the second batch of data from the first batch of data is calculated for each moment based on the estimator. As a result, it is possible to determine the time point at which it can be considered that the tool of the machine tool has worn or chipped. As a result, in the present invention, improvements such as modification of the machining order within a batch can be made. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a block diagram showing the structure of a machine tool inspection system according to a first embodiment of the present invention.

[0024] Figure 2It is a flowchart showing the operation when creating reference data for the omen detection / diagnosis device according to the first embodiment of the present invention.

[0025] Figure 3 (A) of Figure 3 (B) of is a diagram showing an example of the reference data creation result.

[0026] Figure 4 It is a flowchart showing the operation during inspection of the omen detection / diagnosis device according to the first embodiment of the present invention.

[0027] Figure 5 (A) of Figure 5 (B) of is a diagram showing an example of the time series data for reference data and the time series data for inspection.

[0028] Figure 6 It is a diagram showing an example of the time series data of the score.

[0029] Figure 7 It is a flowchart showing the operation when creating reference data for the omen detection / diagnosis device according to the third embodiment of the present invention.

[0030] Figure 8 It is a flowchart showing the processing interval estimation process performed by the reference data creation unit according to the third embodiment of the present invention.

[0031] Figure 9 (A) of Figure 9 (B) of is a diagram explaining the processing interval estimation process performed by the reference data creation unit according to the third embodiment of the present invention.

[0032] Figure 10 It is a diagram explaining the processing interval estimation process performed by the reference data creation unit according to the third embodiment of the present invention.

[0033] Figure 11 It is a flowchart showing the operation during inspection of the omen detection / diagnosis device according to the third embodiment of the present invention.

[0034] Figure 12 It is a block diagram showing the structure of the machine tool inspection system according to the fourth embodiment of the present invention.

[0035] Figure 13 (A) to Figure 13 (D) of is a diagram explaining the influence of the offset component of the batch data on the reference data creation process and the inspection process.

[0036] Figure 14 It is a flowchart showing the operation when creating reference data for the omen detection / diagnosis device according to the fourth embodiment of the present invention.

[0037] Figure 15It is a flowchart showing the offset removal process performed by the data correction execution unit in the fourth embodiment of the present invention.

[0038] Figure 16 It is a diagram showing a display example of a candidate region for removing an offset component in the fourth embodiment of the present invention.

[0039] Figure 17 It is a flowchart showing the operation during the inspection of the omen detection / diagnosis device in the fourth embodiment of the present invention.

[0040] Figure 18 of (A), Figure 18 of (B) are diagrams showing the effects of the fourth embodiment of the present invention.

[0041] Figure 19 It is a block diagram showing the structure of the data correction execution unit in the fifth embodiment of the present invention.

[0042] Figure 20 It is a flowchart showing the operation during the creation of reference data of the omen detection / diagnosis device in the fifth embodiment of the present invention.

[0043] Figure 21 It is a flowchart showing the jitter correction process during the creation of reference data by the data correction execution unit in the fifth embodiment of the present invention.

[0044] Figure 22 It is a flowchart showing the operation during the inspection of the omen detection / diagnosis device in the fifth embodiment of the present invention.

[0045] Figure 23 It is a flowchart showing the jitter correction process during the inspection of the data correction execution unit in the fifth embodiment of the present invention.

[0046] Figure 24 It is a flowchart showing another example of the jitter correction process during the creation of reference data by the data correction execution unit in the fifth embodiment of the present invention.

[0047] Figure 25 It is a flowchart showing another example of the jitter correction process during the inspection of the data correction execution unit in the fifth embodiment of the present invention.

[0048] Figure 26 It is a flowchart showing the operation during the creation of reference data of the omen detection / diagnosis device in the sixth embodiment of the present invention.

[0049] Figure 27 of (A) to Figure 27 of (E) are diagrams showing an example of the average waveform of batch data.

[0050] Figure 28This is a flowchart showing the operation during the inspection of the omen detection / diagnosis device according to the sixth embodiment of the present invention.

[0051] Figure 29 This is a block diagram showing a structural example of a computer that implements the omen detection / diagnosis device according to the first to sixth embodiments.

[0052] [Description of symbols]

[0053] 1, 1a: Omen detection / diagnosis device

[0054] 2: CT

[0055] 3: Display

[0056] 4: Machine tool

[0057] 10: Data acquisition unit

[0058] 11: Data storage unit

[0059] 12: Reference data creation unit

[0060] 13: Inspection execution unit

[0061] 14: Display data generation unit

[0062] 15: Data correction execution unit

[0063] 40: Motor

[0064] 41: Control unit

[0065] 42: CNC

[0066] 150: Cross-correlation calculation unit

[0067] 151: Correction unit Detailed implementation manners

[0068] [First embodiment]

[0069] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. Figure 1 This is a block diagram showing the structure of the machine tool inspection system according to the first embodiment of the present invention. The machine tool inspection system includes: an omen detection / diagnosis device 1; a current transformer (CT) 2, which is a current converter that converts the load current of the motor for driving the workpiece supplied to the machine tool 4 into a magnitude that can be processed by the omen detection / diagnosis device 1; and a display 3 for displaying inspection results and the like obtained by the omen detection / diagnosis device 1.

[0070] The machine tool 4 includes a tool (not shown) for machining a workpiece, a motor 40 for driving the workpiece, a control unit 41 for controlling the motor 40 , and a computer numerical control (CNC) 42 for controlling the entire machine tool.

[0071] The early warning detection / diagnosis device 1 includes: a data acquisition unit 10, which acquires time series data of a load current waveform supplied to a motor 40 of a machine tool 4; a data storage unit 11, which stores the time series data; a reference data creation unit 12, which calculates reference data for performing kernel density estimation for each moment based on batch data for reference data (first batch data) separated from the time series data at the time of reference data creation; an inspection execution unit 13, which calculates a kernel density estimation amount for each moment based on batch data for inspection (second batch data) separated from the time series data at the time of inspection and the reference data, and calculates a score indicating the degree of deviation of the batch data for inspection from the batch data for reference data for each moment based on the estimation amount; and a display data generation unit 14, which generates data to be displayed on the display 3.

[0072] First, the operation of the early warning detection / diagnosis device 1 when creating reference data will be described. Figure 2 This is a flowchart for explaining the operation of the early warning detection / diagnosis device 1 when creating reference data.

[0073] When the machine tool 4 is in a normal state, the user of the early warning detection / diagnosis device 1 instructs the early warning detection / diagnosis device 1 to create reference data.

[0074] The data acquisition unit 10 of the early warning detection / diagnosis device 1, upon receiving an instruction from the user, imports the time series data ( ) of the load current waveform supplied to the motor 40 of the machine tool 4 via the CT 2. Figure 2 Step S100 ) The time series data for reference data imported by the data acquisition unit 10 is stored in the data storage unit 11 .

[0075] The reference data creating unit 12 of the early warning detection / diagnosis device 1 divides the batch data ( Figure 2Step S101). The reference data creation unit 12 divides, for example, the time series data of one batch amount from the machining start time point to the machining end time point as the batch data for reference data. A timing signal that generally represents one batch amount can be obtained from the CNC 42. Alternatively, the reference data creation unit 12 can also determine an interval in which the current value becomes a certain value or more to obtain the cycle length L. For example, if the current value becomes 0.1 A or more, it is determined as the start of one batch, and if the state where the current value is 0.1 A or less continues for 1 second or more, it is determined as the end of one batch, whereby the cycle length L can be obtained. The data acquisition unit 10 and the reference data creation unit 12 repeatedly execute the processes of step S100 and step S101 until the division of the batch data of a pre-specified number of times is completed.

[0076] Next, after the division of the batch data of a pre-specified number of times is completed (in Figure 2 step S102 is "YES"), based on the batch data for reference data, the bandwidth h and the standard deviation σ required for calculating the kernel density estimator to be executed at the time of inspection are calculated Figure 2 step S103). For the sample data x1, sample data x2,..., sample data x n , the function f KDE (x) for calculating the kernel density estimator with the kernel function K(x) and the bandwidth h as parameters can be given by the following formula.

[0077] [Equation 1]

[0078]

[0079] The kernel function K(x) is given by Equation (2).

[0080] [Equation 2]

[0081]

[0082] The bandwidth h is given by Equation (3).

[0083] [Equation 3]

[0084]

[0085] Equations (1) to (3) mean that the function f KDE (x) can be approximated by the sum of the kernel functions K(x) having the same number as the sample data. The interquartile range (IQR) of Equation (3) is the difference between the 75% percentile and the 25% percentile of the samples within the quartile range. min(a, b) is a function that takes the smaller value of a and b.

[0086] The reference data creation unit 12 calculates the standard deviation σ and the bandwidth h at each time of the sample data based on the sample data at times before and after the calculation target time. Here, the time, for example, refers to the elapsed time when the machining start time point is set as time 0. The sample data refers to the current value at a certain time in the batch data. The sample number n is obtained by multiplying the width before and after the calculation target time of the standard deviation σ and the bandwidth h (the number of sample data points at the before and after times) by the number of batch data (the number of batches imported by the data acquisition unit 10).

[0087] Through the above, the creation of the reference data is completed. The calculation results of the reference data (standard deviation σ and bandwidth h) calculated by the reference data creation unit 12 are stored in the data storage unit 11. If the creation of the reference data is implemented using a data set such as Figure 3 (A) as shown, the creation result of the reference data will be as Figure 3 (B) as shown.

[0088] Next, the operation of the omen detection / diagnosis device 1 during inspection will be described. Figure 4 It is a flowchart showing the operation of the omen detection / diagnosis device 1 during inspection.

[0089] The user of the omen detection / diagnosis device 1 instructs the omen detection / diagnosis device 1 to execute the inspection of the machine tool 4.

[0090] When the data acquisition unit 10 of the omen detection / diagnosis device 1 receives an instruction from the user, in the same manner as during the creation of the reference data, it imports the time series data of the load current waveform supplied to the motor 40 of the machine tool 4 via the CT 2 ( Figure 4 step S200). The time series data for inspection imported by the data acquisition unit 10 is stored in the data storage unit 11.

[0091] The inspection execution unit 13 of the omen detection / diagnosis device 1, in the same manner as during the creation of the reference data, divides the inspection batch data from the time series data for inspection stored in the data storage unit 11 according to the preset conditions ( Figure 4 step S201). Then, the inspection execution unit 13 calculates the kernel density estimator f KDE (x) ( Figure 4 step S202) for each time based on the reference data (standard deviation σ and bandwidth h) and the inspection batch data divided in step S201. In the same manner as during the creation of the reference data, the time, for example, refers to the elapsed time when the machining start time point is set as time 0.

[0092] The inspection execution unit 13 substitutes the sample data at the calculation target time in the inspection batch data segmented in step S201 and the reference data (standard deviation σ and bandwidth h) corresponding to the calculation target time into Expression (1) and Expression (2) to calculate the kernel density estimator f at the calculation target time. KDE (x).

[0093] Subsequently, based on the calculated kernel density estimator f KDE (x), the inspection execution unit 13 calculates a score SC representing the degree of deviation of the sample data for inspection with respect to the sample data for reference data at each time Figure 4 (step S203). The score SC is obtained, for example, by the following formula.

[0094] [Equation 4]

[0095] SC = -log(f KDE (x))...(4)

[0096] The higher the score SC, the greater the degree of deviation of the sample data for inspection with respect to the sample data for reference data. If tool wear or the like occurs, the load current changes. Therefore, abnormality detection can be performed based on the level of the score SC. The time series data of the score SC is stored in the data storage unit 11.

[0097] The display data generation unit 14 of the omen detection / diagnosis device 1, for example, correlates the time series data of the score SC with the time series data for inspection imported by the data acquisition unit 10, charts them, and displays them on the display 3 Figure 4 (step S204). The user checks the chart displayed on the display 3 to determine whether there is an abnormality in the machine tool 4 or the like.

[0098] An example of the inspection result based on this embodiment is shown in Figure 5 (A) of Figure 5 (B) of Figure 6 . Here, (A) of Figure 5 is used as the time series data for reference data, and (B) of Figure 5 is used as the time series data for inspection. In the range near 15 seconds indicated by the dashed line in the time series data for inspection, the peak is different from the time series data for reference data. As shown in Figure 6 which shows the calculation result of the score SC based on this embodiment, it can be seen that the value of the score SC becomes high in the range near 15 seconds indicated by the dashed line.

[0099] As described above, in this embodiment, it is possible to determine the time point at which tool wear or a notch is considered to have occurred. As a result, in this embodiment, improvements in applications such as modification of the machining order within a batch can be made.

[0100] [Second Embodiment]

[0101] Next, a second embodiment of the present invention will be described. In this embodiment, the structure of the machine tool inspection system is the same as that of the first embodiment, so the same Figure 1 symbols are also used for description.

[0102] In this embodiment, the data acquisition unit 10 of the omen detection / diagnosis device 1 imports time series data of the load current waveform via the CT 2, and obtains the variety data of the workpiece as the object to be machined from the CNC 42.

[0103] The reference data creation unit 12 of the omen detection / diagnosis device 1 stores the calculated reference data (standard deviation σ and bandwidth h) in correspondence with the variety data acquired by the data acquisition unit 10 at the time of creating the reference data in the data storage unit 11.

[0104] During inspection, the inspection execution unit 13 of the omen detection / diagnosis device 1 acquires the reference data (standard deviation σ and bandwidth h) corresponding to the variety data acquired by the data acquisition unit 10 from the data storage unit 11, and calculates the kernel density estimator f KDE (x) based on the reference data and the batch data segmented during inspection.

[0105] Other operations are the same as those in the first embodiment. Thus, in this embodiment, the present invention can be applied to the machine tool 4 for machining a variety of workpieces, and the abnormality of the machine tool 4 can be detected for each variety of the workpiece.

[0106] [Third Embodiment]

[0107] Next, a third embodiment of the present invention will be described. In this embodiment, the structure of the machine tool inspection system is the same as that of the first and second embodiments, so the same Figure 1 symbols are also used for description.

[0108] In the calculation of the score value during the inspection process, it can be considered that if the interval without a signal is removed and only the values in the interval where the workpiece is actually machined are used for score calculation, the tendency of wear, etc. can be clearly obtained without being affected by the noise in the interval without a signal. Therefore, it is ideal to estimate the interval where the workpiece is actually machined and the interval without a signal where no machining is performed in the batch data.

[0109] Figure 7 is a flowchart showing the operation when creating the reference data of the omen detection / diagnosis device 1 in this embodiment. Figure 7 The processing of steps S100 to S102 is the same as that of the first and second embodiments.

[0110] When machining a workpiece, the load current of the motor 40 of the machine tool 4 increases. The reference data creation unit 12 of the omen detection / diagnosis device 1 detects the section where the load current increases as the section estimated to have been machined, and divides the machining section from the batch data for reference data ( Figure 7 Step S104).

[0111] Figure 8 is a flowchart illustrating the machining section estimation process performed by the reference data creation unit 12. The reference data creation unit 12 calculates a threshold value TH based on the batch data for reference data ( Figure 8 Step S300). The reference data creation unit 12 sets, for example, the current value that is half of the current peak value in the time series data of one batch amount as the threshold value TH. Alternatively, the reference data creation unit 12 may classify the time series data of one batch amount into two categories by, for example, the Otsu binarization method, and set the current value at the midpoint between the centroids of the two categories as the threshold value TH. Alternatively, the reference data creation unit 12 may create a histogram of the current values of the time series data of one batch amount and set the median value as the threshold value TH.

[0112] Subsequently, as shown in (A) of Figure 9 , the reference data creation unit 12 detects the section in the batch data for reference data where the current value is equal to or greater than the threshold value TH as the machining section SP ( Figure 8 Step S301).

[0113] Furthermore, when there is a region in the batch data for reference data where the intervals between the machining sections SP are separated by a time width equal to or greater than a pre-specified time width (when it is "Yes" in Figure 8 Step S302), the process returns to Step S300. In this case, the reference data creation unit 12 calculates the threshold value TH in the region based on the batch data in the region detected in Step S302 (Step S300).

[0114] Then, the reference data creation unit 12 detects the section in the batch data in the detected region where the current value is equal to or greater than the threshold value TH as the machining section SP (Step S301). In the example of (B) of Figure 9 , machining sections SP are newly detected near 30 seconds and near 40 to 45 seconds.

[0115] The processes of Step S300 to Step S302 are repeatedly executed until the region where the intervals between the machining sections SP are separated by a time width equal to or greater than a pre-specified time width disappears. When it is determined to be "No" in Step S302, the detection result as shown in Figure 10 is obtained. The reference data creation unit 12 performs the machining section estimation process as described above for each batch data for reference data.

[0116] Next, for the batch data within the machining section SP, the reference data creation unit 12 calculates the bandwidth h and the standard deviation σ for each machining section SP and each time Figure 7 Step S103a). Through the above, the creation of reference data is completed.

[0117] Figure 11 is a flowchart showing the operation during the inspection of the omen detection / diagnosis device 1 of the present embodiment. Figure 11 The processing of steps S200 and S201 is the same as that of the first and second embodiments. The inspection execution unit 13 of the omen detection / diagnosis device 1 calculates the kernel density estimator f for each time based on the reference data (standard deviation σ and bandwidth h) and the batch data for inspection segmented in step S201 KDE (x)( Figure 11 Step S202a).

[0118] However, in the present embodiment, the reference data (standard deviation σ and bandwidth h) is calculated only within the machining section SP by the reference data creation unit 12. Therefore, the inspection execution unit 13 calculates the kernel density estimator f only for the interval in the batch data for inspection where the elapsed time from the start of the batch data for inspection is the same as that of the machining section SP KDE (x).

[0119] Subsequently, the inspection execution unit 13 calculates a score SC indicating the degree of deviation of the sample data for inspection from the sample data for reference data for each time based on the calculated kernel density estimator f KDE (x) Figure 11 Step S203a). In the present embodiment, the inspection execution unit 13 calculates the score SC only for the interval at the same time position as within the machining section SP.

[0120] Figure 11 The processing of step S204 is the same as that of the first and second embodiments.

[0121] As described above, in the present embodiment, by estimating the machining section SP and calculating the score SC only within the machining section SP, it is possible to detect abnormalities of the machine tool 4 without being affected by the noise in the signal-free section where no machining is performed.

[0122] [Fourth Embodiment]

[0123] Next, a fourth embodiment of the present invention will be described. Figure 12 is a block diagram showing the configuration of the machine tool inspection system according to the fourth embodiment of the present invention. The machine tool inspection system of the present embodiment includes an omen detection / diagnosis device 1a, a CT 2, and a display 3.

[0124] The omen detection / diagnosis device 1a includes a data acquisition unit 10, a data storage unit 11, a reference data creation unit 12, an inspection execution unit 13, a display data generation unit 14, and a data correction execution unit 15.

[0125] If a DC offset component overlaps with the batch data, it will affect the results of the reference data creation process or the inspection process. In the reference data creation process, when using a data group with a small difference in offsets between multiple batch data as shown in Figure 13 (A) of, the result of creating the reference data will be as shown in Figure 13 (B). On the other hand, when using a data group with a large difference in offsets between multiple batch data as shown in Figure 13 (C), the result of creating the reference data will be as shown in Figure 13 (D). If the result of creating the reference data becomes different, the kernel density estimator calculated in the inspection process will also be a different result.

[0126] Even if a data group with a small difference in offsets is used in the reference data creation process, when using data with a small difference in offsets from the batch data used in the reference data creation process and data with a large difference in offsets in the inspection process, the calculation results of the kernel density estimator are different. For the above reasons, it is desirable to appropriately remove the offset component of the batch data.

[0127] Figure 14 is a flowchart showing the operation when creating reference data of the omen detection / diagnosis device 1a of the present embodiment. Figure 14 The processing of steps S100 to S102 and step S104 is the same as that of the third embodiment.

[0128] The data correction execution unit 15 of the omen detection / diagnosis device 1a removes the offset component of the batch data for the reference data segmented in step S101 ( Figure 14 step S105). Figure 15 is a flowchart showing the offset removal process performed by the data correction execution unit 15.

[0129] The data correction execution unit 15 instructs the display data generation unit 14 to display the candidate area CA for removing the offset component in the batch data for the reference data on the display 3 ( Figure 15 step S400). The data correction execution unit 15 detects, in the batch data for the reference data, an area where the current value is equal to or less than a preset offset threshold OTH (OTH > 0) and the state where the current value is equal to or less than the offset threshold OTH continues for a specified time or more as the candidate area CA. The offset threshold OTH is set to a value near 0.

[0130] An example of the display of the candidate area CA displayed on the screen 30 of the display 3 is illustrated in Figure 16 . The user of the omen detection / diagnosis device 1a selects one of the displayed candidate areas CA or designates an arbitrary range.

[0131] The data correction execution unit 15 determines the offset value Ioff based on the batch data within the area designated by the user ( Figure 15 Step S401). For example, the data correction execution unit 15 sets the current value at the point where the slope of the current value in the batch data within the area designated by the user becomes the smallest as the offset value Ioff. Alternatively, for example, the data correction execution unit 15 may also set the mode value of the current values within the area designated by the user as the offset value Ioff.

[0132] The data correction execution unit 15 removes the offset component by subtracting the offset value Ioff from the batch data for reference data ( Figure 15 Step S402). The data correction execution unit 15 performs the above-described offset removal process for each batch data for reference data.

[0133] The reference data creation unit 12 of the present embodiment calculates the bandwidth h and the standard deviation σ for each processing section SP and each moment for the batch data within the processing section SP after the offset removal process ( Figure 14 Step S103b). Through the above, the creation of the reference data is completed.

[0134] Figure 17 is a flowchart showing the operation during the inspection of the omen detection / diagnosis device 1a of the present embodiment. Figure 17 The processing in steps S200 and S201 is the same as that in the first to third embodiments. The data correction execution unit 15 of the omen detection / diagnosis device 1a removes the offset component of the batch data for inspection divided in step S201 ( Figure 17 Step S205). The offset removal process at this time is the same as that at the time of creating the reference data. As the offset value Ioff, the value obtained at the time of creating the reference data may also be used.

[0135] The inspection execution unit 13 of the omen detection / diagnosis device 1a calculates the kernel density estimator f for each moment based on the reference data (standard deviation σ and bandwidth h) and the batch data after the offset removal process in step S205 KDE (x) ( Figure 17 Step S202b). Similar to the third embodiment, the inspection execution unit 13 calculates the kernel density estimator f KDE (x) only for the interval at the same time position as the processing section SP.

[0136] Figure 17The processing of step S203a and step S204 is the same as that of the third embodiment.

[0137] For example, even when, as shown in (A) of Figure 18 , a data group with a large difference in offsets between multiple batch data is used for reference data creation, as long as the offset component is removed by this embodiment, batch data as shown in (B) of Figure 18 can be obtained.

[0138] As described above, in this embodiment, the offset component overlapping with the batch data can be removed, so the influence of the offset component on the result of the reference data creation process or the result of the inspection process can be reduced.

[0139] In addition, in the above description, an example in which the data correction execution unit 15 is applied to the third embodiment has been described, but it can also be applied to the first embodiment and the second embodiment. That is, the processing of Figure 14 step S104 can also be omitted. In this case, for the batch data for reference data after the offset removal process, the bandwidth h and the standard deviation σ are calculated for each moment by the reference data creation unit 12. In addition, the inspection execution unit 13 calculates the kernel density estimator f KDE (x) for each moment based on the reference data (standard deviation σ and bandwidth h) and the batch data for inspection that is segmented in step S201 and has been subjected to the offset removal process in step S205.

[0140] [Fifth Embodiment]

[0141] Next, a fifth embodiment of the present invention will be described. In this embodiment, the structure of the machine tool inspection system is the same as that of the fourth embodiment, so the same Figure 12 symbols are also used for description.

[0142] In the calculation of reference data in the reference data creation process or the calculation of the score value in the inspection process, if jitter occurs in the batch data, it will affect the result of the reference data creation process or the inspection process. Therefore, it is desirable to appropriately correct the jitter.

[0143] Figure 19 is a block diagram showing the structure of the data correction execution unit 15 of this embodiment. The data correction execution unit 15 of this embodiment includes a cross-correlation calculation unit 150 and a correction unit 151.

[0144] Figure 20 is a flowchart for explaining the operation at the time of reference data creation of the omen detection / diagnosis device 1a of this embodiment. Figure 20 The processing of steps S100 to S102 and step S104 of

[0145] Similar to the fourth embodiment, the data correction execution unit 15 in this embodiment removes the offset component of the batch data for the reference data segmented in step S101 ( Figure 20 step S105), and corrects the jitter of the batch data after the offset removal process ( Figure 20 step S106).

[0146] Figure 21 is a flowchart for explaining the jitter correction process of the data correction execution unit 15. The cross-correlation calculation unit 150 of the data correction execution unit 15 calculates the cross-correlation between the processing section SP of one batch of data (hereinafter referred to as representative data) and the block section at the same time position as the processing section SP of the representative data in another batch of data other than the representative data among the batch data for creating reference data segmented by the reference data creation unit 12 a predetermined number of times ( Figure 21 step S501). The cross-correlation calculation unit 150 repeatedly calculates the cross-correlation between the representative data and the block section of another batch of data within a specified time range (for example, TC - window to TC + window) with the start position timing TC of the processing section SP of the representative data as the center (the window is the specified maximum time width).

[0147] After all the cross-correlations within the specified time range are calculated, the correction unit 151 of the data correction execution unit 15 uses the start position timing TC of the processing section SP as the reference position, and obtains the deviation amount of the time position where the cross-correlation becomes the maximum with respect to the reference position ( Figure 21 step S502).

[0148] The cross-correlation calculation unit 150 and the correction unit 151 perform the processing of step S501 and step S502 for each processing section SP of the representative data. When the calculation process of the deviation amount is completed for all the processing sections SP, the correction unit 151 moves the time position of the block section of another batch of data for which the cross-correlation has been calculated by an amount corresponding to the deviation amount calculated for the processing section SP corresponding to the block section, thereby correcting the jitter ( Figure 21 step S503). The correction unit 151 performs the jitter correction process of step S503 for each processing section SP.

[0149] When the correction unit 151 moves the block section of another batch of data in the direction of time delay, for example, it deletes an amount corresponding to the deviation amount of the data behind the end of the block section. Furthermore, since there is a blank part without data in front of the front end of the block section, the correction unit 151 uses the data immediately before the block section to interpolate the blank part corresponding to the deviation amount.

[0150] In addition, when the correction unit 151 moves the block interval of another batch of data in the direction of advancing in time, it deletes an amount corresponding to the deviation amount from the data that is earlier than the front end of the block interval. Further, since there is a blank portion without data after the rear end of the block interval, the correction unit 151 interpolates the blank portion corresponding to the deviation amount using the data immediately following the block interval.

[0151] The cross-correlation calculation unit 150 and the correction unit 151 perform the processing of steps S500 to S503 for each of the other batch data other than the representative data among the batch data for creating reference data divided by the reference data creation unit 12 a predetermined number of times. At the time point when the processing of steps S500 to S503 is completed for all other batch data other than the representative data (in Figure 21 step S504 is "Yes"), the jitter correction process ends. In this way, the correction is performed so as to synchronize the start times of the respective processing intervals SP of the batch data for creating reference data. The batch data after the jitter correction process is stored in the data storage unit 11.

[0152] The reference data creation unit 12 of the present embodiment calculates the bandwidth h and the standard deviation σ for each processing interval SP and each time point for the batch data within the processing interval SP after the offset removal process and then the jitter correction process is performed ( Figure 20 step S103c). Through the above, the reference data creation is completed.

[0153] Figure 22 It is a flowchart showing the operation during the inspection of the omen detection / diagnosis device 1a of the present embodiment. Figure 22 The processing of steps S200 and S201 is the same as that of the fourth embodiment.

[0154] The data correction execution unit 15 of the present embodiment, similarly to the fourth embodiment, removes the offset component of the batch data for inspection divided in step S201 ( Figure 22 step S205), and corrects the jitter of the batch data after the offset removal process ( Figure 22 step S206).

[0155] Figure 23 It is a flowchart showing the jitter correction process during the inspection of the data correction execution unit 15. The cross-correlation calculation unit 150 of the data correction execution unit 15 calculates the cross-correlation between the processing interval SP of the representative data and the block interval at the same time position as the processing interval SP of the representative data in the batch data for inspection after the offset removal process. Figure 23Step S601). The cross-correlation calculation unit 150 repeatedly calculates the cross-correlation between the representative data and the block interval of the batch data for inspection, with the start position timing TC of the processing interval SP of the representative data as the center, within a specified time range (e.g., TC - window to TC + window).

[0156] After all the cross-correlations within the specified time range are calculated, the correction unit 151 of the data correction execution unit 15 uses the start position timing TC of the processing interval SP as the reference position, and obtains the deviation amount of the time position where the cross-correlation becomes the maximum with respect to the reference position ( Figure 23 Step S602).

[0157] The cross-correlation calculation unit 150 and the correction unit 151 perform the processing of Step S601 and Step S602 for each processing interval SP of the representative data. When the calculation processing of the deviation amount is completed for all the processing intervals SP, the correction unit 151 moves the time position of the block interval of the batch data for inspection by an amount corresponding to the deviation amount calculated for the processing interval SP corresponding to the block interval, thereby correcting the jitter. Figure 23 Step S603). The correction unit 151 performs the jitter correction processing of Step S603 for each processing interval SP.

[0158] When the processing of Step S603 is completed for all the processing intervals SP of the representative data, the jitter correction processing ends. In this way, the correction is performed so as to synchronize the start times of each processing interval SP of the representative data and the batch data for inspection. The batch data for inspection after the jitter correction processing is stored in the data storage unit 11.

[0159] The inspection execution unit 13 of the omen detection / diagnosis device 1a calculates the kernel density estimator f for each moment based on the reference data (standard deviation σ and bandwidth h) and the batch data after the jitter correction processing of Step S206. KDE (x)( Figure 22 Step S202c). Similar to the third embodiment, the inspection execution unit 13 calculates the kernel density estimator f KDE (x) only for the interval at the same time position as the processing interval SP.

[0160] Figure 22 The processing of Step S203a and Step S204 is the same as that of the third embodiment.

[0161] As described above, in this embodiment, since the jitter of the batch data can be corrected, the influence of the jitter on the result of the reference data creation processing or the inspection processing can be reduced.

[0162] In addition, as the jitter correction processing, the following processing can also be implemented. Figure 24It is a flowchart showing another example of the jitter correction process during reference data creation. Figure 24 The processing of step S501 and step S502 in Figure 21 is as described. When the deviation calculation process for the first processing section SP of the representative data ends in the correction section 151 of the data correction execution unit 15, the entire time positions of the block sections of the other batch data corresponding to each processing section SP of the representative data are shifted by an amount corresponding to the deviation calculated for the first processing section SP of the representative data, thereby correcting the jitter ( Figure 24 step S505).

[0163] Next, when the deviation calculation process for the second processing section SP of the representative data ends, the entire time positions of the second and subsequent block sections of the other batch data corresponding to the second and subsequent processing sections SP of the representative data are shifted by an amount corresponding to the deviation calculated for the second processing section SP of the representative data, thereby correcting the jitter (step S505).

[0164] Subsequently, when the deviation calculation process for the third processing section SP of the representative data ends, the entire time positions of the third and subsequent block sections of the other batch data corresponding to the third and subsequent processing sections SP of the representative data are shifted by an amount corresponding to the deviation calculated for the third processing section SP of the representative data, thereby correcting the jitter (step S505).

[0165] The cross-correlation calculation unit 150 and the correction unit 151 perform the processing of step S501, step S502, and step S505 as described above for each processing section SP of the representative data. Furthermore, the cross-correlation calculation unit 150 and the correction unit 151 perform the processing of step S500 to step S502 and step S505 for each of the other batch data other than the representative data among the batch data for reference data creation divided by the reference data creation unit 12 a predetermined number of times. At the time point when the processing of step S500 to step S503 is completed for all other batch data other than the representative data (when it is “Yes” in Figure 24 step S506), the jitter correction process during reference data creation ends.

[0166] Figure 25 It is a flowchart showing another example of the jitter correction process during inspection. Figure 25 The processing of step S601 and step S602 in Figure 23As described. When the correction unit 151 of the data correction execution unit 15 finishes the calculation process of the deviation amount for the first processing section SP of the representative data, it moves all the time positions of the block sections of the inspection batch data corresponding to each processing section SP of the representative data by an amount corresponding to the deviation amount calculated for the first processing section SP of the representative data, thereby correcting the jitter ( Figure 25 Step S605).

[0167] Next, when the correction unit 151 finishes the calculation process of the deviation amount for the second processing section SP of the representative data, it moves all the time positions of the second and subsequent block sections of the inspection batch data corresponding to the second and subsequent processing sections SP of the representative data by an amount corresponding to the deviation amount calculated for the second processing section SP of the representative data, thereby correcting the jitter (Step S605).

[0168] Subsequently, when the correction unit 151 finishes the calculation process of the deviation amount for the third processing section SP of the representative data, it moves all the time positions of the third and subsequent block sections of the inspection batch data corresponding to the third and subsequent processing sections SP of the representative data by an amount corresponding to the deviation amount calculated for the third processing section SP of the representative data, thereby correcting the jitter (Step S605).

[0169] The cross-correlation calculation unit 150 and the correction unit 151 perform the processing of Step S601, Step S602, and Step S605 as described above for each processing section SP of the representative data. At the time point when the processing of Step S601, Step S602, and Step S605 is finished for all the processing sections SP of the representative data (when it is "Yes" in Figure 25 Step S600), the jitter correction processing during inspection ends. Figure 24 、 Figure 25 The processing described in is effective when the jitter is large.

[0170] [Sixth Embodiment]

[0171] Next, a sixth embodiment of the present invention will be described. In this embodiment, the structure of the machine tool inspection system is the same as that of the fourth embodiment and the fifth embodiment, so the same Figure 12 reference numerals are also used for description.

[0172] When machining workpieces of multiple varieties using one machine tool 4, it is necessary to hold the reference data of multiple varieties. In the second embodiment, the variety data was obtained from the CNC 42, but in order to obtain the variety data, in many cases, it is also necessary to modify the machine tool 4, which requires time and technology. Therefore, the function of discriminating the workpiece variety based on the batch data becomes useful.

[0173] Figure 26 This is a flowchart showing the operations during the creation of reference data for the omen detection / diagnosis device 1a of this embodiment. Figure 26 The processing of steps S100 to S102 and steps S104 to S106 is the same as that of the fifth embodiment.

[0174] The reference data creation unit 12 of this embodiment calculates the bandwidth h and the standard deviation σ for each processing section SP and each moment for the batch data for reference data within the processing section SP that has undergone offset removal processing and then jitter correction processing. Figure 26 Step S103d). At this time, the reference data creation unit 12 obtains the average waveform of the multiple batch data for reference data used in the calculation of the bandwidth h and the standard deviation σ, and the average cycle length L of the multiple batch data for reference data. ave as the characteristic quantities of the batch data for reference data.

[0175] Since the reference data creation unit 12 can obtain the cycle length of the time series data of one batch amount by acquiring a timing signal indicating one batch amount from the CNC 42, the average cycle length L can be calculated by obtaining the cycle length for each of the multiple batch data. ave . Alternatively, the reference data creation unit 12 can also determine an interval where the current value is a certain value or more to obtain the cycle length L. For example, if the current value becomes 0.1 A or more, it is determined as the start of one batch, and if the state where the current value is 0.1 A or less continues for 1 second or more, it is determined as the end of one batch, whereby the cycle length L can be obtained. Then, the reference data creation unit 12 stores the reference data including the bandwidth h, the standard deviation σ, the average waveform of the batch data, the average cycle length L ave and the offset value Ioff determined by the data correction execution unit 15 in correspondence with the workpiece type data in the data storage unit 11.

[0176] The reference data creation unit 12 performs the reference data creation process as described above for each type of workpiece. The workpiece type data during reference data creation is manually input by the user of the omen detection / diagnosis device 1a for each type of workpiece. Thus, the reference data creation unit 12 can perform the reference data creation process for each type of workpiece.

[0177] Examples of the average waveforms of the batch data for workpiece types A to E are shown respectively in Figure 27 (A) to Figure 27 (E) of

[0178] Figure 28 This is a flowchart showing the operations during the inspection of the omen detection / diagnosis device 1a of this embodiment. Figure 28The processing of step S200 and step S201 is the same as that of the fifth embodiment.

[0179] The inspection execution unit 13 of this embodiment obtains the cycle length L of the inspection batch data segmented in step S201 ( Figure 28 step S207). The inspection execution unit 13 can obtain the cycle length L of the time series data of one batch volume by obtaining a timing signal representing one batch volume from the CNC 42. Alternatively, the inspection execution unit 13 can also determine an interval in which the current value is equal to or greater than a certain value to obtain the cycle length L. For example, if the current value becomes 0.1 A or more, it is determined as the start of one batch, and if the state where the current value is 0.1 A or less continues for 1 second or more, it is determined as the end of one batch, thereby obtaining the cycle length L.

[0180] Subsequently, the inspection execution unit 13 compares the cycle length L obtained in step S207 with the average cycle length L included in the reference data for each variety of the workpiece ave and determines how many varieties of workpieces satisfy the specified variety determination conditions ( ave step S208). For example, when the cycle length L falls within ±5% of the average cycle length L, the inspection execution unit 13 regards the variety of the workpiece corresponding to the reference data including this average cycle length L Figure 28 as the variety that satisfies the variety determination conditions and sets it as the first variety candidate. ave ave ave When there is only one first variety candidate, the inspection execution unit 13 determines the first variety candidate as the variety of the workpiece during inspection (

[0181] step S209). In addition, when there is no first variety candidate, the inspection execution unit 13 sets it as not being able to perform variety determination ( Figure 28 step S210). Figure 28 step S210).

[0182] In Figure 27 example (A) to Figure 27 example (E), varieties A, B, D, and E become the first variety candidates.

[0183] When the inspection execution unit 13 determines that there are two or more first variety candidates, the data correction execution unit 15 of this embodiment corrects the jitter of the inspection batch data segmented in step S201. Figure 28Step S211). At this time, the data correction execution unit 15 generates batch data subjected to jitter correction processing for each of the first-round variety candidates. For example, when varieties A, B, D, and E are the first-round variety candidates, the data correction execution unit 15 generates inspection batch data obtained by subjecting the average waveform of the usage batch data of variety A to jitter correction processing, inspection batch data obtained by subjecting the average waveform of the usage batch data of variety B to jitter correction processing, inspection batch data obtained by subjecting the average waveform of the usage batch data of variety D to jitter correction processing, and inspection batch data obtained by subjecting the average waveform of the usage batch data of variety E to jitter correction processing.

[0184] The inspection execution unit 13 calculates the correlation coefficient or cross-correlation of the average waveforms of the inspection batch data subjected to jitter correction processing and the batch data included in the reference data of the first-round variety candidates for each of the first-round variety candidates ( Figure 28 Step S212), and determines how many workpieces of the varieties that satisfy the specified variety determination conditions with respect to the correlation coefficient or cross-correlation ( Figure 28 Step S213).

[0185] The inspection execution unit 13 sets the variety candidates with a correlation coefficient or cross-correlation equal to or higher than the reference value as the second-round variety candidates that satisfy the variety determination conditions. In the Figure 27 example of (A) to Figure 27 example of (E), variety A is below the reference value, and varieties B, D, and E are above the reference value. Thus, varieties B, D, and E become the second-round variety candidates.

[0186] When there is only one second-round variety candidate, the inspection execution unit 13 determines the second-round variety candidate as the variety of the workpiece during inspection (Step S209). In addition, when there is no second-round variety candidate, the inspection execution unit 13 sets it as impossible to perform variety determination (Step S210).

[0187] When the inspection execution unit 13 determines that there are two or more second-round variety candidates, the data correction execution unit 15 of this embodiment generates batch data with the offset component removed for each of the second-round variety candidates by subtracting the offset value Ioff from the inspection batch data ( Figure 28 Step S214).

[0188] At this time, the data correction execution unit 15 subtracts the offset value Ioff included in the reference data of the variety candidate from the batch data for inspection obtained by performing jitter correction processing on the average waveform of the usage batch data of the second variety candidate. Therefore, the data correction execution unit 15 subtracts the offset value Ioff of variety B from the batch data for inspection obtained by performing jitter correction processing on the average waveform of the usage batch data of variety B, subtracts the offset value Ioff of variety D from the batch data for inspection obtained by performing jitter correction processing on the average waveform of the usage batch data of variety D, and subtracts the offset value Ioff of variety E from the batch data for inspection obtained by performing jitter correction processing on the average waveform of the usage batch data of variety E. In addition, when the position of the offset point does not become a second variety candidate due to reasons such as insufficient cycle length, the variety is excluded from the second variety candidate.

[0189] Next, the inspection execution unit 13 calculates the kernel density estimator f for each of the second variety candidates and each moment based on the reference data (standard deviation σ and bandwidth h) and the batch data after the offset removal process in step S214. KDE (x)( Figure 28 Step S215). Similarly to the third embodiment, the inspection execution unit 13 calculates the kernel density estimator f only for the interval at the same time position as the processing interval SP. KDE (x). The inspection execution unit 13 can substitute the batch data for which the offset removal process has been performed using the offset value Ioff of variety B and the reference data of variety B into equations (1) and (2) to calculate the kernel density estimator f. KDE (x). The same applies to variety D and variety E.

[0190] Subsequently, the inspection execution unit 13 calculates the score SC for each of the second variety candidates and each moment based on the calculated kernel density estimator f. KDE (x), for example, using equation (4), and calculates the average value of the scores SC of one batch quantity for each of the second variety candidates ( Figure 28 Step S216). For example, the average value of the score SC of variety B is 2.11, the average value of the score SC of variety D is 5.83, and the average value of the score SC of variety E is 6.82.

[0191] The inspection execution unit 13 determines how many variety candidates have an average value of the score SC below a certain value ( Figure 28 Step S217). When there is no second variety candidate with an average value of the score SC below a certain value, the inspection execution unit 13 sets it as impossible to perform variety determination (step S210).

[0192] When there are one or more candidate varieties whose average score SC is below a certain value, the inspection execution unit 13 determines the variety of the workpiece at the time of inspection as the second candidate variety with the smallest average score SC, and uses the score SC calculated for the candidate variety as the inspection result. Figure 28 (Step S218). In the above example, variety B is determined as the variety of the workpiece at the time of inspection.

[0193] In addition, when there is one candidate variety in the first candidate variety or the second candidate variety, the process can be performed based on the determination result of the variety.

[0194] That is, the data correction execution unit 15 subtracts the offset value Ioff included in the reference data of the candidate variety from the inspection batch data obtained by performing the jitter correction process on the average waveform of the use batch data of the candidate variety determined as the variety of the workpiece at the time of inspection, thereby removing the offset component. Figure 28 (Step S219). However, when there is one candidate variety in the first candidate variety, the data correction execution unit 15 performs the offset removal process after performing the jitter correction process on the inspection batch data using the average waveform of the batch data of the candidate variety.

[0195] Next, the inspection execution unit 13 substitutes the inspection batch data after the offset removal process and the reference data of the candidate variety determined as the variety of the workpiece at the time of inspection into Equation (1) and Equation (2), and calculates the kernel density estimator f KDE (x) Figure 28 (Step S220). Subsequently, the inspection execution unit 13 can calculate the score SC for each moment based on the calculated kernel density estimator f KDE (x), for example, using Equation (4). Figure 28 (Step S221).

[0196] When the variety of the workpiece at the time of inspection is determined through the process of Step S218, the calculation process of the kernel density estimator f KDE (x) and the score SC has been completed. Therefore, the score SC calculated for the candidate variety determined as the variety of the workpiece at the time of inspection as described above can be used as the inspection result.

[0197] Figure 28 The process of Step S204 is the same as that of the first to fifth embodiments.

[0198] In this embodiment, it is not necessary to obtain the product type data from the machine tool 4, and only the load current needs to be obtained. Additionally, even when the same type of workpiece is being machined using the same machine tool 4, due to the influence of the machining environment, accidental errors, poor machining, etc., there will be deviations in the motor current distribution. In this embodiment, by using kernel density estimators or correlations between the reference data and the target data, it is possible to determine the product type while allowing for these deviations.

[0199] In addition, in the above description, an example of applying this embodiment to the fifth embodiment has been described, but the processing of step S104 Figure 26 can also be omitted.

[0200] The premonition detection / diagnosis device 1 and the premonition detection / diagnosis device 1a described in the first to sixth embodiments can be implemented by a computer including a central processing unit (CPU), a storage device, and an interface, as well as a program for controlling these hardware resources. An example of the structure of the computer is illustrated in Figure 29 .

[0201] The computer includes a CPU 200, a storage device 201, and an interface device (interface, I / F) 202. A display 3, a machine tool 4, etc. are connected to the I / F 202. In such a computer, the program for implementing the premonition detection / diagnosis method of the present invention is stored in the storage device 201. The CPU 200 executes the processing described in the first to sixth embodiments in accordance with the program stored in the storage device 201. Additionally, at least a part of the premonition detection / diagnosis device 1 and the premonition detection / diagnosis device 1a can also be implemented using hardware.

[0202] [Industrial Applicability]

[0203] The present invention can be applied to the technology for detecting abnormalities in machine tools.

Claims

1. A warning detection / diagnosis device, characterized in that: include: a data acquisition unit configured to acquire time series data of a load current supplied to a motor of a machine tool; A data storage unit configured to store the time series data; a reference data creation unit configured to calculate reference data for kernel density estimation at each time point based on a first batch of data segmented from the time series data at the time of reference data creation; as well as The inspection execution unit is configured to calculate a kernel density estimator for each time point based on second batch data separated from the time series data during the inspection and the reference data, and calculate a score indicating a degree of deviation of the second batch data from the first batch data for each time point based on the kernel density estimator.

2. The omen detection / diagnosis device according to claim 1, characterized in that: The data correction execution unit is further included, the data correction execution unit being configured to detect an offset value for a region in the batch data where a current value is less than an offset threshold and the state of the current value being less than the offset threshold continues for more than a predetermined time, and to subtract or add the offset value from the batch data.

3. The omen detection / diagnosis device according to claim 2, characterized in that: The data correction execution unit corrects the plurality of first batch data so that their time positions coincide with each other, and corrects the second batch data so that their time positions coincide with those of the first batch data.

4. The omen detection / diagnosis device according to claim 3, characterized in that: The reference data creation unit stores the feature quantities of the plurality of first batch data and the reference data in association with the type data of the workpiece input by the user in the data storage unit. The inspection execution unit determines the type of workpieces during inspection based on the feature amount of the first batch data and the second batch data, and calculates the kernel density estimation amount based on the reference data corresponding to the type of workpieces during inspection and the second batch data.

5. The omen detection / diagnosis device according to claim 4, characterized in that: The characteristic quantities of the first batch of data are the average waveform and average cycle length of the batch of data, The inspection execution unit performs a first variety determination by comparing the average cycle length with the cycle length of the second batch data. When there are two or more variety candidates in the first variety determination, a second variety determination is performed based on the correlation coefficient or cross-correlation between the average waveform and the second batch data. When there are two or more variety candidates in the second variety determination, the kernel density estimator is calculated for each variety candidate remaining in the second variety determination, and the variety of the workpiece during inspection is determined based on the kernel density estimator.

6. The omen detection / diagnosis device according to any one of claims 1 to 5, characterized in that: The reference data creating unit detects a processing section in which workpiece processing is estimated to be performed in the first batch data. The inspection execution unit calculates the kernel density estimate only for the section at the same time position as the processing section in the second batch data.

7. The omen detection / diagnosis device according to claim 1, characterized in that: The data acquisition unit imports the time series data and acquires the type data of the workpiece from the machine tool. The reference data creation unit stores the calculated reference data in the data storage unit in correspondence with the product data acquired by the data acquisition unit when the reference data is created. The inspection execution unit acquires the reference data corresponding to the product data acquired by the data acquisition unit from the data storage unit during the inspection, and calculates a kernel density estimate based on the reference data and the second batch data.

8. A method for early warning detection / diagnosis, characterized in that: include: The first step is to obtain time series data of the load current supplied to the motor of the machine tool; The second step is to calculate the benchmark data for kernel density estimation at each moment based on the first batch of data segmented from the time series data when the benchmark data is created; The third step is to calculate a kernel density estimator for each moment based on the second batch of data segmented from the time series data during the inspection and the reference data; as well as In the fourth step, based on the kernel density estimator, a score indicating the degree of deviation of the second batch of data relative to the first batch of data is calculated for each moment.

Images